First year Algebra:
eigenvalues and eigenvectors
Let A be a n n matrix.
is an eigenvalue for A if there exists a vector u 0 such that
Au u.
If such a vector u exists, it is said to be an eigenvector associated with the
eigenvalue .
(I) How to find the eigenvalues and eigenvectors
Suppose that is an eigenvalue for A .
x1
Then there exists a non-zero vector u ... such that Au u or,
x
n
equivalently, such that ( A I n )u 0 , where I n is the n n identity matrix.
Since ( A I n )u 0 has a non-trivial solution u , the matrix ( A I n ) is not
invertible, i.e. A I n 0.
Note that A I n is a polynomial in , so you get the eigenvalues of A by
finding the roots of A I n .
Say 0 is such an eigenvalue. In order to find the eigenvectors associated
with 0 , you have to solve the system Au 0 u for x1 ,..., xn .
(II) Example: two by two matrices
3 2
.
Let A
2 1
To find the eigenvalues of A , form the matrix A I 2 , find its determinant
A I 2 and solve the equation A I 2 0 :
2
3 2
1 0 3
.
A I 2
1
2 1
0 1 2
Now,
A I 2 (3 )(1 ) (2)(2) 2 2 1 ,
so that the eigenvalues of A are the roots of 2 2 1 ( 1) 2 , i.e. A has a
repeated eigenvalue: 0 1 .
Now we have to solve the system Au 0 u . Here 0 1 , so that
Au 0 u
Au u ,
which yields
x
3 2 x1
1
2 1 x2
x2
or, equivalently,
2 2 x1 0
.
2 2 x2 0
x
A vector u 1 is therefore an eigenvector associated with the eigenvalue
x2
0 1 if and only if its coordinates satisfy
2 x1
2 x1
2 x2 0
2 x2 0
,
i.e. if and only if x1 x2 .
Hence the eigenvectors associated with the eigenvalue 0 1 are of the
1
form u , where is a real number.
1
(III)
Example: three by three matrices
1
3 1 1
3
Let A 7 5 1 . Then A I 3 7
5
6 6 2
6
6
A I 3 (3 )
5
6
1
1 and
2
1
7
1
7 5
3 12 16 .
2 6 2 6
6
In general, it is not easy to find the roots of a polynomial of degree 3. We try
some simple values and hope we will be lucky. For 2 , you get
A I 3 (2) 3 12(2) 16 0 ,
so that 2 is a root of A I 3 .
Hence ( 2) is a factor of A I 3 and A I 3 can be written in the form
A I 3 ( 2)(a2 b c) a3 2 (b 2a) (c 2b) 2c.
We now have
A I 3 3 12 16 a3 2 (b 2a) (c 2b) 2c,
so that a, b and c must satisfy
a 1
b 2a 0
,
c 2b 12
2c 16
which yields
a 1
b2.
c8
We are now in position to find all the factors of A I 3 :
A I 3 ( 2)(2 2 8) ( 2) 2 ( 4).
A therefore has two eigenvalues: 0 2 and 1 4 .
To get the eigenvectors associated with 0 2 , we solve ( A 2I 3 )u 0 :
1 1 1 x1 0
x1
( A 2 I 3 )u 7 7 1 x2 0 7 x1
6 6 0 x 0
6 x
1
3
x2
7 x2
6 x2
x3 0
x x2
x3 0 1
.
x3 0
0
Hence the eigenvectors associated with the eigenvalue 0 2 are of the
1
form 1 .
0
To get the eigenvectors associated with 1 4 , we solve ( A 4I 3 )u 0 :
7 1 1 x1 0
7 x1
( A 2 I 3 )u 7 1 1 x2 0 7 x1
6 6 6 x 0
6 x
1
3
x2
x2
6 x2
x3 0
x 0
x3 0 1
.
x 2 x3
6 x3 0
Hence the eigenvectors associated with the eigenvalue 1 4 are of the form
0
1 .
1
(III)
Exercises
Find the eigenvalues and the eigenvectors of the matrix A , where
(i)
0 2
A
3 5
(ii)
0 5
A
0 1
(iii)
1 3 3
A 3 5 3
6 6 4